PTSD symptoms, pain catastrophizing, and pain outcomes after acute orthopedic injury

Abstract

Background

Posttraumatic stress disorder (PTSD) is associated with greater incidence of chronic pain. Pain catastrophizing often accounts for this association. Less is known about these relationships during the acute phase (1–2 months) following orthopedic traumatic injuries. We sought to understand which orthopedic traumatic injury-related PTSD symptoms were associated with acute pain and physical dysfunction and whether pain catastrophizing accounted for these associations.

Methods

This secondary analysis used baseline data from a multisite randomized controlled trial of an intervention for individuals with heightened pain catastrophizing or pain anxiety following acute orthopedic injury. We used partial correlations to examine associations between PTSD symptom clusters (re-experiencing, avoidance, negative alterations in cognitions and mood, and hyperarousal) and pain outcomes (pain intensity and physical dysfunction) controlling for pain catastrophizing. We used hierarchical regressions to evaluate unique associations between PTSD clusters and pain outcomes. In exploratory analysis, we examined the indirect effects of PTSD symptoms on pain outcomes through catastrophizing.

Results

Hierarchical linear regressions indicated that hyperarousal was uniquely associated with greater pain intensity with activity (β = 0.39, P < .001, ΔR² = 0.06) and physical dysfunction (β = 0.22, P = 0.04, ΔR² = 0.02). PTSD symptoms were still associated with pain with activity even with pain catastrophizing included in the models, and catastrophizing did not have a significant indirect effect on the relationship between PTSD and physical dysfunction (b = 0.06, SE_Boot = 0.04, 95% CI_Boot = [−0.003, 0.14]). Pain catastrophizing did largely account for the association between re-experiencing, avoidance, and negative alterations in cognitions and mood symptoms and pain at rest.

Conclusions

Pain catastrophizing interventions may be best suited for limiting the impact of PTSD symptoms on pain at rest, but catastrophizing alone may not fully explain the relationship between PTSD symptoms and physical dysfunction after acute orthopedic injury. To prevent the negative association of PTSD symptoms, especially hyperarousal, on physical outcomes in acute pain populations, interventions may require more than solely targeting pain catastrophizing.

Clinical trials registration

NCT04973696.

Introduction

Posttraumatic stress disorder (PTSD) symptoms are associated with nervous system dysregulation leading to a range of health problems, including poor general health,^1–3 cardiovascular disease,^4–6 dysregulated immune responses,⁷ and heightened incidences of chronic pain conditions.^8–11 Even a single trauma exposure is enough to increase the risk for health problems, including pain.¹² PTSD is associated with sympathetic nervous system over-activation and central sensitization, which have been shown to be associated with unfavorable pain outcomes, including higher pain intensity and reduced functioning.^13,14 Often, acute orthopedic injuries (fractures, ruptures, dislocations) are traumatic in and of themselves, and symptoms of PTSD after acute orthopedic injury predict higher pain severity and greater impairment over time.^11,15,16

However, PTSD is a heterogeneous condition, and it remains unclear which symptoms contribute to physical dysfunction in those who experience acute orthopedic injury. PTSD represents a complex array of symptoms characterized by 4 clusters: (1) re-experiencing of the traumatic event (eg, experiencing intrusive thoughts of the event, nightmares, flashbacks, or intense emotional or physiological reactions to reminders of the event), (2) avoidance (eg, of internal or external reminders of the traumatic event), (3) alterations in negative cognitions and mood (eg, negative emotions such as sadness or shame, negative thoughts about the self or the world, feelings of interpersonal disconnection), and (4) hyperarousal (eg, hypervigilance, anger behaviors, sleep disturbance, engagement in risky behaviors). Broadly speaking, re-experiencing and hyperarousal represent heightened responsiveness to threat, whereas avoidance and alterations in negative cognitions and mood signify withdrawal from threat.¹⁷ In a large-scale study of trauma survivors, hyperarousal and avoidance demonstrated independent associations with functional impairment for individuals with PTSD,¹⁸ suggesting that these broad patterns of approach and withdrawal from threat can both lead to dysfunction independently.

Similarly, approach and withdrawal from threat have emerged as predictors of pain after injury. Hyperarousal and emotional numbing symptoms (a combination of avoidance and negative alterations in cognitions and mood) have been shown to cluster together and predict the development of disability after whiplash injury.¹⁹ Six months after burn injury, hyperarousal and emotional numbing also predicted pain interference above the other symptom clusters and pain intensity.²⁰ In chronic pain after injury, hyperarousal symptoms uniquely predicted pain intensity, while avoidance/numbing were associated with activity avoidance.²¹ However, we do not know which PTSD criteria drive the relationship with pain intensity and physical dysfunction after acute orthopedic injury. Regarding the clinical impact on recovery from injury, it is possible that re-experiencing and hyperarousal symptoms could lead to more of the anxious activation that is tied to heightened pain intensity,²² while avoidance and negative alterations in cognitions and mood could lead patients to become more detached from physical activity and their lives broadly, which could slow recovery. Acute pain (ie, < 3 months) is an important window of time to prevent the development of chronic pain, as psychological factors in this timeframe are some of the most robust predictors of future pain and functional interference.^23–26 Understanding which PTSD symptom clusters are associated with pain and functioning outcomes is important for specifying treatment targets in the acute injury phase.

Additionally, little is understood about factors accounting for the relationships between PTSD symptoms, pain intensity, and physical dysfunction among individuals with acute orthopedic injuries. In the chronic pain literature, pain catastrophizing has consistently been shown to account for the association between PTSD symptoms and pain outcomes, including the development of chronic pain after injury.^27–30 While fear of pain and pain anxiety are sometimes implicated in these associations,³¹ pain catastrophizing is the factor that most consistently accounts for the relationship between PTSD and pain.^28,29 Catastrophizing represents a potential shared mechanism implicated in maintenance models of PTSD and in the fear-avoidance model of chronic pain, which posits that catastrophic reactions to pain lead to greater dysfunction over time.^28,32 These previous findings suggest that targeting pain catastrophizing in treatment would be sufficient to reduce the impact of PTSD on pain. However, it is unclear whether this is true for acute pain, and whether pain catastrophizing is equally important across PTSD symptom clusters. Pain catastrophizing is a cognitive construct and thus may map onto PTSD criteria that are more cognitive in nature, such as re-experiencing and negative alterations in cognition and mood, as opposed to avoidance and hyperarousal, but catastrophizing also can have downstream effects on avoidance of activity and physiological arousal.^22,33 Thus, whether catastrophizing can fully account for the effects of PTSD on acute orthopedic injury outcomes is unclear. Additionally, while many studies have examined associations between PTSD, catastrophizing, and pain intensity, far fewer studies have examined physical dysfunction as an outcome, despite its importance as an indicator of the extent to which an individual can live meaningfully despite pain.³⁴ Physical dysfunction represents difficulty completing any physical task, including activities of daily living and physically active chores or exercise.³⁵ It is crucial to understand the relationship between PTSD and physical dysfunction early after orthopedic injury so that we can develop targeted interventions that can prevent long-term pain interference.^24,36

In the present analyses, we examined the associations between PTSD symptoms, pain catastrophizing, pain intensity, and physical dysfunction among a sample of adults with acute orthopedic injuries. We hypothesized that the avoidance and hyperarousal criteria of PTSD would be uniquely associated with both pain intensity and physical dysfunction. We additionally hypothesized that the associations between PTSD criteria—re-experiencing, avoidance, negative cognitions/emotions, and hyperarousal—with both pain intensity and physical dysfunction would be eliminated when pain catastrophizing was taken into consideration. In exploratory analyses, we hypothesized that pain catastrophizing would account for the association between PTSD and pain outcomes through a significant indirect effect of PTSD symptoms (full scale and subscales) on pain intensity and physical dysfunction through pain catastrophizing.

Methods

Participants

This is a secondary analysis of data from a study evaluating a mind-body intervention delivered to patients with orthopedic injury who exhibit heightened emotional reactivity to pain (either pain catastrophizing or pain anxiety) and are thus at higher risk for transitioning from acute to chronic pain. Full details of the study and full CONSORT diagram are reported elsewhere.³⁷ Participants were recruited from 4 geographically diverse Level 1 Trauma Centers at different sites around the United States. Potential participants were identified through screening of the medical record by a trained research assistant. Referring providers were then requested to conduct a warm hand-off to introduce the study to the participant. The research assistant then administered the screening questions to evaluate eligibility. Participants were eligible to participate if they: (1) were outpatient adults at 1 of the 4 Level 1 Trauma Centers participating in this study; (2) were recovering from acute musculoskeletal injury (eg, fracture, dislocation, rupture—ie, injured internal tissue, such as a tendon rupture) that occurred 1–2 months prior to contact about this study; (3) obtained Pain Catastrophizing Scale (PCS) ≥ 20 or Pain Anxiety Symptoms Scale Short Version (PASS-20) ≥ 40; (4) owned a smartphone, laptop, or computer; (5) had access to the internet; (6) were willing to comply with the study protocol; (7) were free of psychotropics or stable (ie, no changes in type of medication or dose) for > 6 weeks; (8) were cleared by the orthopedic surgeon for study participation, (9) were able to meaningfully participate (eg, speak English); and (10) scored at least 8/10 on the Short Portable Mental Status Questionnaire for patients above 65 years of age or those staff identified as having cognitive impairment. Exclusion criteria included: (1) serious comorbidity expected to worsen in the next 3 months; (2) serious untreated mental illness; (3) current suicidal ideation with plan; (4) other concurrent severe injuries that could impact participation (such as TBI), (5) current substance use disorder within the past 6 months; (6) current litigation or under worker’s compensation; (7) surgical complications (infection, need for repeat surgery); (8) practice of yoga/meditation, or other mind body techniques, once per week for 45 min or more within the last 3 months, and (9) self-reported pregnancy (as required by Institutional Review Board). See CONSORT diagram in the main outcome paper for full details of study flow.³⁷ Briefly, 2818 medical records were screened, 936 were excluded because of the above exclusion criteria, withdrawing or declining consent, or other logistical reasons prior to baseline, and overall 351 patients were selected for inclusion with PASS-20 or PCS scores above the required cutoffs. All participants were cleared to participate by their surgeon.

Procedures

This is a secondary analysis of the baseline data from a study examining the Toolkit for Optimal Recovery (TOR) intervention for individuals with heightened pain catastrophizing or pain anxiety following acute orthopedic injury.³⁸ The TOR intervention consisted of 4 weekly live video individual visits of 45 min each with a PhD-level psychologist with expertise in pain management and mind body skills. The study was approved by the Institutional Review Boards at all study sites. Participants were screened for eligibility, and those who met criteria for study participation were invited to participate. All participants were invited to review and discuss consent with trained study staff prior to signing consent and participating in study procedures. Participants then completed baseline questionnaires on a tablet prior to being randomized to the active intervention or control. Participants were included in the present analyses if they completed baseline questionnaires and the first session of the study intervention.

Data collection

At baseline, participants were requested to provide demographic information (age, gender, race/ethnicity, education, employment, income, marital status, current psychotropic/pain medication intake, comorbid medical conditions, history of mental health conditions) and were administered a battery of measures.

PTSD symptoms as related to the acute orthopedic injury were assessed with the Posttraumatic Stress Disorder Checklist-Civilian version (PCL-C),³⁹ a 20-item self-report questionnaire answered on a 5-point Likert scale (0 = “Not at all” to 4 = “Extremely”), eg, “In the past month, how much were you bothered by: Repeated disturbing, and unwanted memories of the stressful experience?”. The questionnaire assesses 3 domains based on DSM-IV PTSD criteria, but for the purposes of the present study, we scored 4 subscales based on DSM-5 PTSD criteria: Re-experiencing, avoidance, negative alterations in cognition and mood, and hyperarousal. In the instructions for the measure, symptoms were anchored to orthopedic injury with the following instructions, “Below is a list of problems and complaints that people sometimes have in response to stressful life experiences, such as suffering an acute orthopedic injury. Please read each one carefully, then select one answer to indicate how much you have been bothered by that problem in the last month. The stressful experience refers to your recent orthopedic injury.”

Pain catastrophizing was assessed using the Pain Catastrophizing Scale (PCS),⁴⁰ a 13-item self-report questionnaire measured on a 5-point Likert scale (0 = “Not at all” to 4 = “Always”). The PCS assesses catastrophizing across 3 domains—magnification (eg, “It’s awful and I feel that it overwhelms me”), helplessness (eg, “There’s nothing I can do to reduce the intensity of the pain”), and rumination (eg, “I worry all the time about whether the pain will end”).

Regarding pain outcomes, pain intensity when active was measured by the Numeric Rating Scale,⁴¹ a reliable pain-intensity measure on an 11-point Likert scale of 0 “No pain” to 10 “Worst imaginable pain.” Pain was assessed at rest and during activity. Physical dysfunction was assessed with the Short Musculoskeletal Function Assessment—Dysfunction Index (SMFA),⁴² a 46-item subscale of the SMFA. The Dysfunction Index is scored on a 5-point Likert scale on which participants rate their physical ability on each item, eg, “How difficult is it for you to get in or out of a low chair.” Higher scores indicate greater physical dysfunction. We used only the Dysfunction Index to focus specifically on perceived physical ability, whereas the total score additionally assesses the extent to which an individual is bothered by perceived limitations.

Measures selected were identified as reliable and valid^43–47 and have been used in prior research testing the main study intervention’s feasibility.⁴⁸

Data analytic plan

All analyses were completed in SPSS Version 28. All measures were normally distributed with skewness and kurtosis within the values of −1 and 1. We generated descriptive statistics to characterize variables and used Pearson correlations to evaluate bivariate associations between the PCL-C total score, the subscales, pain intensity, and physical dysfunction. We also included partial correlations controlling for pain catastrophizing to demonstrate attenuation of associations between PTSD symptoms and outcomes when catastrophizing was taken into consideration. To examine whether there were unique associations between PTSD symptom clusters and outcomes (pain intensity and physical dysfunction), we conducted 2 hierarchical regressions in which the re-experiencing, avoidance, negative alterations in cognition and mood, and hyperarousal subscales were entered separately on each step with either pain intensity or physical dysfunction as the outcome. We examined standardized beta coefficients to compare the strength of the effect of each individual independent variable on the dependent variable. Finally, we conducted exploratory analyses to examine the indirect effect of PCL-C on pain intensity and physical dysfunction through pain catastrophizing (2 mediation models). We utilized the PROCESS macro in SPSS⁴⁹ with bootstrapping (5000 samples). Analyses were run with and without controlling for demographic variables (age, sex, race, and income) that have been shown to be associated with either PTSD symptoms or pain intensity/recovery from injury.^50,51

Results

Participant characteristics

Participants included 181 adults (M = 44.16, SD = 16.50) across the 4 study sites (Site A N = 63, Site B N = 44, Site C N = 44, Site D N = 30). The majority of the sample (68.5%) identified as female (N = 124) and White (76.2%). Full demographic information can be found in Table 1. The sample reported experiencing minimal pain at rest (M = 3.78, SD = 2.39), though reported pain with activity was moderate (M = 6.02, SD = 2.51). Participants endorsed experiencing moderate levels of pain catastrophizing (M = 23.17, SD = 10.78), which was expected due to the inclusion criteria for the present study requiring PCS scores ≥20 and/or pain anxiety scores ≥40. Patients also endorsed moderate physical dysfunction (M = 46.67, SD = 15.02) as measured by the SMFA—Dysfunction Index. Most participants in the sample presented with fractures (n = 160, 88.4%) with the remaining experiencing dislocation (n = 3, 1.7%), rupture of tendon or ligament (n = 3, 1.7%), multiple injuries (n = 11, 6.1%), and unspecified injury (n = 4, 2.2%).

Table 1.

Participant characteristics.

Participant characteristics (N = 181)	Participant, no. (%)
Age	Mean = 44.16 (SD = 16.50)
Sex
Female	124 (68.5%)
Male	57 (31.5%)
Race
American Indian/Alaskan Native	1 (0.6%)
Asian	9 (5.0%)
Black/African American	14 (7.7%)
White	138 (76.2%)
More than 1 race	11 (6.1%)
Choose not to answer	8 (4.4%)
Ethnicity
Hispanic or Latino/a	25 (13.8%)
Not Hispanic or Latino/a	151 (83.4%)
Choose not to answer	5 (2.8%)
Marital status
Single, never married	70 (38.7%)
Married	70 (38.7%)
Living with significant other	8 (4.4%)
Separated/divorced	23 (12.7%)
Widowed	9 (5.0%)
Choose not to answer	1 (0.6%)
Education
Less than high school (<12 years)	7 (3.9%)
Completed high school or GED (12 years)	42 (23.2%)
Some college/associates degree (16 years)	47 (26.0%)
Completed 4 years of college (16 years)	47 (26.0%)
Graduate/professional degree (> 16 years)	37 (20.4%)
Choose not to answer	1 (0.6%)
Employment
Employed full-time	89 (49.2%)
Employed part-time	24 (13.3%)
Keeping house/housemaker	3 (1.7%)
Going to school full or part-time	6 (3.3%)
Retired	17 (9.4%)
Unemployed	26 (14.4%)
Other	10 (5.5%)
Choose not to answer	6 (3.3%)
Income
Less than $10 000	26 (14.4%)
$10 000 to less than $15 000	10 (5.5%)
$15 000 to less than $20 000	4 (2.2%)
$20 000 to less than $25 000	10 (5.5%)
$25 000 to less than $35 000	14 (7.7%)
$35 000 to less than $50 000	24 (13.3%)
$50 000 to less than $75 000	18 (9.9%)
$75 000 or more	51 (28.2%)
Choose not to answer	24 (13.3%)

Correlations between PTSD symptoms, pain intensity, and physical dysfunction

PTSD symptoms (PCL total scores and subscales) were significantly correlated with pain intensity and physical dysfunction. These relationships mostly remained significant for pain intensity when active and physical dysfunction, even after controlling for pain catastrophizing, with the exception of the association between re-experiencing symptoms (Criterion B) and pain intensity when active (Table 2). The association between PTSD symptoms and pain intensity at rest were significantly attenuated for re-experiencing, avoidance, and negative alterations in cognitions or mood. Controlling for demographic variables did not change the results.

Table 2.

. Correlations and partial correlations controlling for pain catastrophizing among the clinical sample (N = 181).

	Mean (SD)	1.	2.	3.	4.	5.	6.	7.	8.	9.
1. PCL total	40.01 (15.85)	–
2. PCL_B	11.20 (5.64)	0.92*** (0.89***)	–
3. PCL_C	4.73 (2.39)	0.81*** (0.79***)	0.77*** (0.74***)	–
4. PCL_D	11.63 (5.18)	0.89*** (0.87***)	0.71*** (0.65***)	0.62*** (0.56***)	–
5. PCL_E	12.43 (4.76)	0.88*** (0.86***)	0.71*** (0.66***)	0.61*** (0.56***)	0.72*** (0.67***)	–
6. Pain at rest	3.78 (2.39)	0.38*** (0.22**)	0.32*** (0.14)	0.28*** (0.15*)	0.29*** (0.13)	0.42*** (0.31***)	–
7. Pain when active	6.02 (2.51)	0.36*** (0.22**)	0.28*** (0.11)	0.31*** (0.20**)	0.30*** (0.16*)	0.40*** (0.30***)	0.71*** (0.64***)	–
8. Physical dysfunction	46.67 (15.02)	0.46*** (0.37***)	0.40*** (0.30***)	0.33** (0.24**)	0.43*** (0.35***)	0.44*** (0.36***)	0.40*** (0.30***)	0.37*** (0.28)***	–
9. Pain catastrophizing	23.17 (10.78)	0.45***	0.47***	0.34***	0.39***	0.37***	0.45***	0.40***	0.33***	–

^* P < .05,

^** P < .01,

^*** P < .001.

Values in parentheses indicate coefficients from partial correlations controlling for pain catastrophizing. PCL_B to PCL_E, subscales for PTSD criteria. B: reexperiencing, C: avoidance, D: negative alterations in cognition/mood, E: hyperarousal.

Relationship between PTSD subscales and pain intensity

Separate models were run for pain intensity at rest and when active. Results were largely the same. Step 1 of the hierarchical linear regression model testing re-experiencing (Criterion B) as a predictor of pain intensity was statistically significant. In this model, re-experiencing was associated with pain intensity. In Step 2, where avoidance (Criterion C) was added to the model, the model was statistically significant, but the addition of avoidance did not account for a significant increase in R². Further, in the pain when active model, neither re-experiencing nor avoidance contributed significantly to this model. In Step 3, where negative alterations in cognition and mood (Criterion D) were added to the model, the model was statistically significant, but the addition of Criterion D did not account for a significant increase in R². In this model, neither Criterion B, C, or D contributed significantly to the model (Ps > .05). In Step 4, where hyperarousal (Criterion E) was added to the model, the model was significant and accounted for a significant increase in R². In this model, only hyperarousal was associated with pain intensity, such that increased hyperarousal symptoms were associated with greater pain intensity (Table 3). Results remained the same when controlling for demographic variables.

Table 3.

Hierarchical regression models examining the association between PTSD and pain intensity and physical dysfunction.

Model	Step and predictor	β	p	ΔR2
Pain at rest	Step 1			0.11***
	PCL_B	0.33	<0.001
	Step 2			0.002
	PCL_B	0.27	0.02
	PCL_C	0.08	0.49
	Step 3			0.01
	PCL_B	0.21	0.11
	PCL_C	0.06	0.60
	PCL_D	0.10	0.31
	Step 4			0.07***
	PCL_B	0.07	0.61
	PCL_C	0.02	0.84
	PCL_D	−0.07	0.53
	PCL_E	0.41	<0.001
Pain when active	Step 1			0.08***
	PCL_B	0.28	<0.001
	Step 2			0.02
	PCL_B	0.11	0.32
	PCL_C	0.22	0.05
	Step 3			0.01
	PCL_B	0.02	0.89
	PCL_C	0.19	0.10
	PCL_D	0.17	0.10
	Step 4			0.06***
	PCL_B	−0.12	0.35
	PCL_C	0.16	0.16
	PCL_D	0.001	0.99
	PCL_E	0.39	<0.001
Physical dysfunction	Step 1			0.16***
	PCL_B	0.40	<0.001
	Step 2			0.001
	PCL_B	0.36	0.001
	PCL_C	0.05	0.62
	Step 3			0.04**
	PCL_B	0.19	0.11
	PCL_C	0.001	0.99
	PCL_D	0.29	0.003
	Step 4			0.02*
	PCL_B	0.11	0.36
	PCL_C	−0.02	0.87
	PCL_D	0.20	0.06
	PCL_E	0.22	0.04

^* P < .05,

^** P < .01,

^*** P < .001.

PCL_B to PCL_E = subscales for PTSD criteria. B: re-experiencing, C: avoidance, D: negative alterations in cognition/mood, E: hyperarousal.

Relationship between PTSD subscales and physical dysfunction

Step 1 of the hierarchical linear regression model testing re-experiencing (Criterion B) as a predictor of physical dysfunction was statistically significant. In this model, re-experiencing was associated with physical dysfunction, such that increased re-experiencing symptoms were associated with decreased physical dysfunction. In Step 2, where avoidance (Criterion C) was added to the model, the model was statistically significant, but the addition of avoidance did not account for a significant increase in R². Further, only re-experiencing contributed significantly to the model, and avoidance was not significant. In Step 3, where negative alterations in cognition and mood (Criterion D) was added to the model, the model was statistically significant and accounted for a significant increase in R². In this model, only Criterion D was associated with physical dysfunction, such that increased negative alterations in cognition and mood symptoms were associated with decreased physical dysfunction. In Step 4, where hyperarousal (Criterion E) was added to the model, the model was significant and accounted for a significant increase in R². In this model, only hyperarousal was associated with physical dysfunction, such that increased hyperarousal symptoms were associated with decreased physical dysfunction (Table 3). Results largely remained the same when controlling for demographic variables, except for an additional significant unique association between Criterion D and physical dysfunction (β = 0.22, P = .04) when all PTSD subscales and demographic variables were included in the model.

Indirect effect of pain catastrophizing on the relationship between PTSD symptoms and pain intensity at rest

Total effects, direct effects, and indirect effects of all models are presented in Table 4. PTSD symptoms (full-scale PCL-C) were associated with pain catastrophizing (path a: b = 0.31, SE = 0.05, 95% CI = [0.22, 0.40], P < .001). The total effect of PTSD on pain intensity at rest was statistically significant (path c: b = 0.06, SE = 0.01, 95% CI = [0.04, 0.08], P < .001). Increased pain catastrophizing was associated with increased pain intensity at rest (path b: b = 0.08, SE = 0.02, 95% CI = [0.05, 0.11], P < .001). PTSD symptoms accounted for a significant amount of the variance in pain intensity at rest (R² = 0.14), and the inclusion of pain catastrophizing added a significant amount of additional variance (change in R² = 0.10, P < .001). There was still a statistically significant direct effect of PTSD symptoms on pain intensity at rest after accounting for catastrophizing (path c’: b = 0.03, SE = 0.01, 95% CI = [0.01, 0.05], P = .004). but the indirect effect of PTSD symptoms on pain intensity at rest through pain catastrophizing was also significant (path a × b: b = 0.02, SE_Boot = 0.01, 95% CI_Boot = [0.01, 0.04]). Results were similar for the Criterion C and E subscales, which were still significantly associated with pain intensity at rest when pain catastrophizing was in the model. By contrast, there was no longer a significant association between pain intensity at rest and Criterion B (b = 0.06, SE = 0.03, 95% CI = [−0.002, 0.12], P = .06) or D (b = 0.06, SE = 0.03, 95% CI = [−0.01, 0.12], P = .07) when catastrophizing was accounted for. However, the associations between all PCL subscales and pain intensity at rest were significant with demographic variables in the model.

Table 4.

Total, direct, and indirect effects of PTSD symptoms on pain intensity and physical dysfunction through pain catastrophizing.

	B	Standard error	P value	Lower confidence interval	Upper confidence interval
Pain at rest
PCL total
Total effect	0.06	0.01	<.001	0.04	0.08
Direct effect (ie, accounting for catastrophizing)	0.03	0.01	.004	0.01	0.05
Indirect effect	0.02	0.01		0.01	0.04
PCL criterion B
Total effect	0.14	0.03	<.001	0.08	0.20
Direct effect	0.06	0.03	0.06	−0.002	0.12
Indirect effect	0.08	0.04		0.11	0.26
PCL criterion C
Total effect	0.28	0.07	<.001	0.14	0.43
Direct effect	0.15	0.07	.04	0.01	0.29
Indirect effect	0.14	0.04		0.07	0.21
PCL criterion D
Total effect	0.13	0.03	<.001	0.07	0.20
Direct effect	0.06	0.03	.07	−0.01	0.12
Indirect effect	0.07	0.02		0.04	0.11
PCL criterion E
Total effect	0.21	0.03	<.001	0.14	0.28
Direct effect	0.15	0.03	<.001	0.08	0.21
Indirect effect	0.06	0.02		0.04	0.10
Pain when active
PCL total
Total effect	0.06	0.01	<.001	0.04	0.08
Direct effect	0.04	0.01	.003	0.01	0.06
Indirect effect	0.02	0.01		0.01	0.03
PCL criterion B
Total effect	0.12	0.03	<.001	0.06	0.19
Direct effect	0.05	0.03	.13	−0.02	0.12
Indirect effect	0.07	0.02		0.04	0.11
PCL criterion C
Total effect	0.32	0.07	<.001	0.17	0.47
Direct effect	0.20	0.08	.01	0.06	0.35
Indirect effect	0.12	0.03		0.06	0.19
PCL criterion D
Total effect	0.14	0.03	<.001	0.08	0.21
Direct effect	0.08	0.04	.03	0.01	0.15
Indirect effect	0.06	0.02		0.03	0.10
PCL criterion E
Total effect	0.21	0.04	<.001	0.14	0.28
Direct effect	0.16	0.04	<.001	0.08	0.23
Indirect effect	0.06	0.02		0.03	0.09
Physical dysfunction
PCL total
Total effect	0.44	0.06	<.001	0.31	0.56
Direct effect	0.37	0.07	<.001	0.24	0.51
Indirect effect	0.06	0.04		−0.003	0.14
PCL criterion B
Total effect	1.06	0.18	<.001	0.70	1.42
Direct effect	0.84	0.20	<.001	0.43	1.24
Indirect effect	0.23	0.10		0.03	0.43
PCL criterion C
Total effect	2.07	0.44	<.001	1.19	2.94
Direct effect	1.54	0.46	.001	0.64	2.45
Indirect effect	0.52	0.20		0.18	0.96
PCL criterion D
Total effect	1.24	0.20	<.001	0.86	1.63
Direct effect	1.03	0.21	<.001	0.62	1.44
Indirect effect	0.21	0.09		0.04	0.40
PCL criterion E
Total effect	1.37	0.21	<.001	0.95	1.79
Direct effect	1.14	0.22	<.001	0.70	1.59
Indirect effect	0.23	0.09		0.05	0.42

Indirect effect of pain catastrophizing on the relationship between PTSD symptoms and pain intensity when active

PTSD symptoms (full-scale PCL-C) were associated with pain catastrophizing (path a: b = 0.31, SE = 0.05, 95% CI = [0.22, 0.40], P < .001). Increased pain catastrophizing was associated with increased pain intensity when active (path b: b = 0.07, SE = 0.02, 95% CI = [0.04, 0.10], P < .001). There was a statistically significant total effect of PTSD symptoms on pain intensity when active (path c: b = 0.06, SE = 0.01, 95% CI = [0.04, 0.08], P < .001). PTSD symptoms accounted for a significant amount of the variance in pain intensity when active (R² = 0.13), and the inclusion of pain catastrophizing added a significant amount of additional variance (change in R² = 0.07, P < .001). The direct effect of PTSD on pain intensity when active, after accounting for catastrophizing, was still statistically significant (path c’: b = 0.04, SE = 0.01, 95% CI = [0.01, 0.06], P = .003), but the indirect effect of PTSD symptoms on pain intensity when active through pain catastrophizing was also significant (path a × b: b = 0.02, SE_Boot = 0.01, 95% CI_Boot = [0.01, 0.03]) (Figure 1). Results were similar for the Criterion C, D, and E subscales all still significantly associated with pain intensity when active when pain catastrophizing was in the model. The exception was for Criterion B, which was no longer significantly associated with pain intensity when active when catastrophizing was accounted for (b = 0.05, SE = 0.03, 95% CI = [−0.02, 0.12], P = .13). Results remained the same when controlling for demographic variables.

Figure 1.

Indirect effect of PTSD symptoms (full-scale PCL -C) on pain intensity when active through pain catastrophizing.

Indirect effect of pain catastrophizing on the relationship between PTSD symptoms and physical dysfunction

PTSD symptoms (full-scale PCL-C) were associated with pain catastrophizing (path a: b = 0.31, SE = 0.05, 95% CI = [0.22, 0.40], P < .001). Increased pain catastrophizing was associated with decreased physical dysfunction (path b: b = 0.21, SE = 0.10, 95% CI = [0.01, 0.41], P = 0.04). The total effect of PTSD on decreased physical dysfunction was statistically significant (path c: b = 0.44, SE = 0.06, 95% CI = [0.31, 0.56], P < .001). PTSD symptoms accounted for a significant amount of the variance in physical dysfunction (R² = 0.21), and the inclusion of pain catastrophizing added a significant amount of additional variance (change in R² = 0.02, P = .04). Nevertheless, there was a statistically significant direct effect of PTSD symptoms on decreased physical dysfunction even when accounting for pain catastrophizing (path c’: b = 0.37, SE = 0.07, 95% CI = [0.24, 0.51], P < .001), and the indirect effect of PTSD symptoms on decreased physical dysfunction through pain catastrophizing was not significant (path a × b: b = 0.06, SE_Boot = 0.04, 95% CI_Boot = [−0.003, 0.14]). Results were similar for all criterion subscales still significantly associated with physical dysfunction when pain catastrophizing was in the model, despite a significant indirect effect. Results remained the same when controlling for demographic variables.

Discussion

In this sample of patients with acute orthopedic injury, we demonstrated associations between pain outcomes (pain intensity and physical dysfunction) and PTSD symptoms anchored to acute orthopedic injury as the traumatic event. This is consistent with a wealth of literature demonstrating an association between acute PTSD symptoms and non-acute pain severity.^52,53 We also reported on the unique associations between PTSD symptom clusters and pain intensity and physical dysfunction outcomes. All symptom clusters were associated with pain intensity and physical dysfunction, but only hyperarousal was uniquely associated with each outcome. Pain catastrophizing partially accounted for these associations and added a significant amount of variance explained, especially for pain intensity at rest. However, with the exception of the re-experiencing symptoms, catastrophizing was not sufficient to fully explain the relationship between PTSD symptoms and pain with physical activity/physical dysfunction.

Our finding of a unique association between hyperarousal and pain intensity/physical dysfunction replicates and extends prior findings. Hyperarousal has previously been shown to predict pain interference after burn injury²⁰ and to uniquely predict pain intensity and interference following whiplash injury.^19,21 We demonstrated the uniquely important role that hyperarousal symptoms play for individuals with acute orthopedic injury. Hyperarousal symptoms, including symptoms such as hypervigilance, elevated anger behaviors, and sleep difficulties, are associated with sympathetic nervous system over-activation, including increased skin conductance/heart rate.^54,55 Such activation has been posited to play a key role in the development and maintenance of chronic pain.⁵⁶ Clinicians treating acute orthopedic injury may benefit from assessing PTSD symptoms early in recovery, providing psychoeducation on the relationship between PTSD nervous system dysregulation and pain, and integrating techniques that specifically target PTSD hyperarousal symptoms, including Dialectical Behavior Therapy distress tolerance skills, exposure, or techniques from Skills Training in Affective and Interpersonal Regulation⁵⁷ concurrently with pain management skills, including those that target pain catastrophizing. One avenue could be providing a conceptualization of the orthopedic injury as a trauma to the patient and then tailoring pain management skills used to reduce hyperarousal (eg, deep breathing, progressive muscle relaxation) to the memory of the traumatic injury.

Previous research has consistently found that catastrophizing fully accounts for the association between PTSD symptoms and chronic pain.^27–30,58 This was consistent with our findings for pain at rest, which makes sense given that perseverative thinking processes such as catastrophizing are most prone to occurring during times with fewer distractions.⁵⁹ Catastrophizing interventions may be best suited for attenuating these relationships. Our findings for pain with activity and physical dysfunction suggest that the association between acute symptoms of PTSD after orthopedic injury and acute pain/physical dysfunction cannot fully be explained by pain catastrophizing for most of the PTSD criteria. Of note, many of our pain treatments are linked to the fear-avoidance model of chronic pain,⁶⁰ which posits a central role of catastrophizing in the prediction of physical dysfunction. Existing pain treatments, including for acute pain, are designed to target catastrophizing through a range of techniques, such as cognitive restructuring and other mind-body skills.^35,48 However, our findings suggest that addressing PTSD symptoms, especially those related to hyperarousal, may be particularly beneficial in conjunction with pain management techniques at the acute injury phase when physical function is the target of intervention.

Limitations

There are several limitations to the present study. Participants were predominantly White and non-Hispanic. PTSD symptoms were self-report only using the PCL-C which captures DSM-IV criteria, and there was no assessment of prior trauma. Finally, this was a cross-sectional study, so mediation models were exploratory, and the direction of associations cannot be confirmed. Further, due to our selection criteria of elevated pain catastrophizing or pain anxiety, we cannot generalize these findings to those with below-clinical levels on these variables. It is possible that due to the range restriction, we reduced the ability to find significant associations between catastrophizing and PTSD symptoms, pain intensity, and physical dysfunction. Our results may also not generalize to others who were excluded from the present sample, including patients who already use mind-body techniques and those with more serious mental illness or substance use. Future research should replicate these findings in a more diverse sample using an interview assessment of DSM-5 PTSD symptoms and a prospective design. Future studies would benefit from targeted recruitment of patients from diverse racial, ethnic, and cultural populations so that we may better understand whether the present findings generalize to all patients experiencing orthopedic injury. Finally, the present study lacked information on the type of event that caused the orthopedic injury. Future work should evaluate whether the event (eg, motor vehicle accident, fall) moderates associations between PTSD symptoms and pain outcomes, as some events are more likely to cause PTSD than others.

Conclusions

PTSD symptoms are important to assess early after orthopedic injury. Hyperarousal symptoms may be particularly important to consider early after acute orthopedic injury due to their unique associations with pain intensity when active and with physical dysfunction. While pain catastrophizing attenuated some of these effects, it could not fully account for most of these associations. Our findings suggest that pain management interventions delivered acutely after orthopedic injury may benefit from the addition of skills to cope with hyperarousal symptoms of PTSD. Addressing PTSD alongside pain and physical dysfunction early after an acute orthopedic injury may prevent both chronic pain and PTSD.

Funding

This study was funded by a National Center for Complementary and Integrative Health (U01AT010462 to A.-M.V.). A.-M.V. was additionally funded by NCCIH (K24AT011760. K.M. was funded by the National Institute on Aging (3R61AG081034-01S1). T.P. was funded by NCCIH (K23AT012363).

Conflicts of interest: None declared.